Cerebras Systems Discloses Q1 2026 Financials and 750 MW Infrastructure Agreement With OpenAI

Cerebras Systems Inc. (NASDAQ: CBRS) reported its financial results for the first quarter of fiscal year 2026 today, balancing significant top-line expansion against ongoing infrastructure buildouts. The hardware and cloud services developer posted GAAP quarterly revenue of $193.4 million, representing a 94% increase compared to the $99.5 million generated in the first quarter of 2025.

Revenue Distribution and Margin Profiles

The company's revenue growth was driven by sharp acceleration in its cloud and services division, indicating a shifting business mix.

Cloud and Services Revenue: GAAP cloud and other services revenue grew 178% year-over-year to $82.8 million, up from $29.8 million in the prior year period.

Hardware Revenue: GAAP hardware revenue totaled $110.6 million, representing a 59% increase over the $60.7 million reported in Q1 2025.

Operational Losses: Cerebras recorded a GAAP loss from operations of $15.0 million and a final GAAP net loss of $14.0 million, narrowing from an operating loss of $28.5 million a year ago.

Gross Margins: GAAP gross margin was reported at 45% overall, with cloud and services operating at a 49% gross margin compared to 41% for hardware shipments.

Cerebras closed the quarter with $3.3 billion in cash, cash equivalents, restricted cash, and short-term investments.

Scale of the OpenAI Infrastructure Commitment

Alongside its earnings release, Cerebras formalized the parameters of a multi-year commercial agreement with OpenAI valued at more than $20 billion. Under the contract, OpenAI has committed to deploying up to 750 megawatts (MW) of Cerebras inference compute capacity over the next several years.

The infrastructure partnership aims to address processing latency constraints for generative models. The companies jointly introduced Codex-Spark, an interactive development model engineered for coding tasks that handles an execution threshold above 1,000 tokens per second. To fund early-stage deployment logistics, OpenAI initially extended a $1 billion working capital loan to Cerebras in January 2026.

Amazon Web Services Hybrid Cloud Architecture

Cerebras also disclosed a multi-year partnership with Amazon Web Services (AWS) to launch a disaggregated cloud inference deployment strategy. The processing configuration splits active model workloads across specialized architectures to achieve greater resource efficiency.

Under this setup, AWS Trainium 3 chips will manage the initial prefill operations, while Cerebras CS-3 wafer-scale systems will handle the high-speed decode execution phase. The integrated offering is designed to scale low-latency inference delivery for enterprise and AI-native applications operating across AWS regions.

Capital Structure and Acquisition Liquidity

The operational scaling follows a major restructuring of the company's capital allocation framework. Cerebras secured $6.4 billion in gross proceeds through its initial public offering in the second quarter, which stands as the largest semiconductor IPO in financial history.

This equity injection followed a $1 billion Series H pre-IPO round completed in February. To expand its pipeline of data center acquisitions, Cerebras closed an additional $850 million revolving credit facility in April from a syndicate of investment banks.

The expanded compute pool is currently hosting enterprise customer trials of Kimi K2.6, a trillion-parameter open-weight model independently verified by Artificial Analysis to approach processing speeds of 1,000 tokens per second, alongside image-understanding evaluations of Google DeepMind's Gemma 4 31B architecture.